From Global Governance to Nationalism: The Future of AI

The 2025 Paris AI Summit has revealed fundamental disagreements over AI governance with many states prioritising AI sovereignty and market dominance over human-rights-based global governance. This has intensified geopolitical and economic rivalries between states and reduced regulatory guardrails against the potential misuse of AI for autocratic purposes by governments and corporations.

• The EU’s human-rights-centric regulations clash with the innovation-driven deregulation in AI governance promoted by the US. Meanwhile, economic powerhouses such as India, South Korea, and Brazil are establishing their own positions. Autocracies such as China, Russia, Iran, the UAE, and Saudi Arabia, with their state-led model of AI governance, are integrating AI with military and surveillance capabilities.

• Geoeconomic rivalries are likely to result in a re-shuffling of AI supply chains and trade barriers between the US, China, the EU, and other AI powers, impeding innovation and creating monopolies. Such barriers make low-income countries in the Global South highly dependent on states that lead in AI development, further reinforcing global inequalities. Countries with permissive regulatory environments are likely to attract more business but at the cost of increased risk of AI misuse.

• As states and powerful technology-developing companies prioritise their narrow interests over broader ethical considerations and public safety, the space for civil society in AI governance is shrinking, which may exacerbate concerns over privacy, employment, and social justice and human rights.

Policy Implications

The EU must maintain its lead in AI governance by enabling innovation with ethics, building sustainable and scalable domestic AI data centres and infrastructure, securing financial support for start-ups, and strengthening academia–industry linkages. Incentivising cooperation with emerging AI economies to align regulatory frameworks and influence global standards is of utmost importance.

Divided Agendas at the Paris AI Summit

The race to dominate global artificial intelligence (AI) governance has accelerated. Over the past two years, major AI summits focused on global collaboration. However, the 2025 Paris AI Summit, the third edition of an annual event bringing together government, industry, and civil society experts to explore AI advancement and develop common policies, marked a turning point. It exposed a stark divide, as the prevailing multi-stakeholder, human-rights-centred model favoured by the European Union (EU) was challenged by representatives of the US government, who emphasised national security and economic growth.

Is AI Reshaping the Global Order?

Global AI governance debates are occurring amid rapid tech advances and strong investment interests. As numerous countries develop their own AI strategies, they also recognise the need for global coordination, as seen in the 2023 UK and 2024 Seoul AI Summit declarations. The Paris AI Summit aimed for greater inclusivity, hosting over 1,000 participants from 100+ countries (including significant Global South representation) compared to the UK Summit’s 100 participants from 30 predominantly AI-advanced countries. In Paris, major powers demonstrated their willingness to compete assertively. While 62 countries signalled their commitment to global best practices by signing the Paris AI Summit's Statement on Inclusive and Sustainable AI (Élysée 2025), the US and the UK declined to sign. Given that the two countries account for approximately 40 per cent of global AI funding, their rejection of the statement severely undermines efforts towards global AI governance (Mukherjee 2024). The event underscored a shift by the US and the UK from leading global AI safety discussions to adopting more inward-focused positions, prioritising national security interests and domestic legislation. Meanwhile, the EU signalled a move towards deregulation, and China, by launching its AI Safety Institute just before the summit, expressed readiness to engage in global collaboration on AI safety and governance.

Countries such as the US, the UK, Israel, Singapore, the UAE, Saudi Arabia, China, and Qatar prioritise business innovation and economic growth through light regulation and industry-led initiatives. By contrast, Japan, South Korea, Germany, and France have adopted a more cautious approach, emphasising comprehensive rules to mitigate risks. Yet another group of countries – India, Brazil, Canada, and Australia – aim to balance rapid AI deployment with responsible implementation by integrating ethical oversight and adaptive regulation.

Meanwhile, the world has reached a critical turning point as governments increasingly adopt AI to drive economic development and technological advancement. Countries across the Global South are looking for AI-driven solutions for national development. For instance, China’s adaptable AI applications that support customisation for regional needs, such as DeepSeek-R1, are used in Africa for education and agriculture. These technologies support indigenous-language development and reduce reliance on English-centric tools.

Across regions, countries form strategic collaborations to advance their AI interests. BRICS positions itself as a counterbalance to the Global North’s tech dominance. Advocating for a multipolar digital world, the bloc claims to also represent technologically disadvantaged countries. Brazil’s 2025 BRICS presidency highlights this by prioritising South–South cooperation in six areas, following G20 initiatives, which, alongside global health, trade, climate change, and security-architecture reforms, also include AI governance. In fact, AI development will be central to the BRICS bloc’s strategy of fostering knowledge transfer among emerging economies. Russia, for instance, faces challenges in AI development due to limited computing power and sanctions, despite successes in transportation and medical platforms. To overcome these hurdles, Russia promotes industrial collaboration and international partnerships (UNCTAD 2024). Thus, in 2024, Putin announced a BRICS-led AI alliance, with the Russian Sberbank financing joint AI research with China (Lee 2025). Although an active BRICS member, South Africa has adopted EU-like regulations through its National AI Policy Framework, thus addressing responsible AI development amid research capacity challenges (Allen 2024). Across Latin America, Brazil, Argentina, and Mexico are drafting AI regulations, though many of these appear to lack precision and enforcement mechanisms. Brazil is emerging as a regional leader through initiatives such as the Ministerial Summit on AI Ethics and the Cartagena Declaration (Levy Daniel 2025).

As tensions between the US and China over AI supremacy persist, African countries try to remain neutral. However, the recent deregulatory turn of the US could create a vacuum that allows China to exert greater influence in shaping international AI policies.

Thus, whereas some countries are leading in both development and regulation, others focus on rapidly advancing their technological capabilities. Meanwhile, a large group of countries is struggling to maintain relevance. For a better understanding of this evolving landscape of AI players, we outline five categories of countries, each reflecting different levels of influence, regulatory approaches, and strategic positioning within the global AI ecosystem.

Global AI Leaders

The world’s preeminent economic and tech powers, the US and China, dominate AI development, while the US and the EU seek to do the same in AI governance. The US leads in AI innovation, with strong private-sector involvement and increasingly tight export controls on advanced AI technologies. The EU emphasises regulation, exemplified by the 2024 EU AI Act, positioning itself as the foremost advocate for ethical AI deployment and digital rights protection. This leadership role was further demonstrated when the OECD, a close EU partner, launched the first global AI reporting framework for companies at the Paris AI Summit. Amongst other things, the OECD advanced the Hiroshima AI Process (HAIP), initiated during Japan’s 2023 G7 presidency, which establishes a voluntary reporting framework to promote transparency and accountability in advanced AI development. Meanwhile, China systematically expands AI capabilities through coordinated state-led initiatives, integrating AI into its sophisticated surveillance infrastructure and using it to extend geopolitical influence (notably through technology exports to the MENA region and its involvement in Zimbabwe’s surveillance systems).

Rapidly Emerging AI Hubs

Industrialised and technologically advanced economies such as India, the UK, Canada, Japan, and South Korea strive to gain competitive advantages in the global AI landscape through strategic policy incentives and targeted technological investments. India has offered technical assistance and capacity-building to South and Southeast Asian countries. Through training in digital public infrastructure (DPI) and its future use in AI, India has set up regional knowledge hubs for AI (for example, in healthcare). Moreover, the South Asian country initiates collaborative projects on AI through its IndiaAI mission. It also aspires to be a global AI player, from chairing the international initiative on responsible and human-centric AI called the Global Partnership for Artificial Intelligence (GPAI) to co-hosting the Paris AI Summit and hosting the fourth edition of the summit in 2026. The UK is attracting global talent and capital with adaptive post-Brexit AI-friendly policies, exemplified by the Alan Turing Institute and increased AI research funding. Canada, too, has established itself as a centre of excellence in AI research, emphasising ethics and safety through initiatives such as its Pan-Canadian AI Strategy and the pioneering work of research centres including the Vector Institute and Mila. Japan and South Korea lead in AI-driven automation and robotics. Developments in Japan are particularly noteworthy: recently, the country pivoted from strengthening its regulatory framework to a more hybrid AI approach, blending European pro-regulation principles with elements of the US-led tech ecosystem model to balance innovation, appropriate regulation, transparency, and ethics (AI Policy Study Group 2025). This shift reflects a growing awareness of AI-specific risks, current methodological limitations in assessing the safety of advanced AI models, regulatory adjustments under Donald Trump’s new US administration, and the global race for AI innovation, standards, and infrastructure control. Furthermore, the loss of the parliamentary majority of Japan’s ruling Liberal Democratic Party in the October 2024 election has created unprecedented political fragmentation within the Diet, complicating the implementation of major legislative reforms. The recent formation of SB OpenAI Japan, a 50–50 joint venture between OpenAI and SoftBank, exemplifies Japan’s hybrid AI approach (Mok and Tong 2025). Over 500 Japanese companies across various industries are involved in developing AI applications tailored to their needs, boosting opportunities for AI application start-ups and helping to create a sustainable AI ecosystem in Japan. Additionally, the collaboration extends beyond Japan: SoftBank and OpenAI are both participating in the USD 500 billion Stargate Project. Thus, Japan maintains its technological sovereignty while engaging in major international initiatives.

Regional AI Influencers

Regional AI leaders including Singapore, Australia, the UAE, and Saudi Arabia are shaping AI governance within their respective neighbourhoods while tailoring strategies to their economic and strategic needs. Singapore has firmly established itself as a regional leader in AI governance and digital infrastructure development with its National AI Strategy and Model AI Governance Framework, serving as influential blueprints for other countries in Southeast Asia. Australia emphasises AI safety, national security, and regulation through specialised government bodies such as the National AI Centre. The UAE and Saudi Arabia integrate AI into their efforts towards economic transformation, driving smart city initiatives and global investments in AI start-ups and research institutions. They are attempting to position themselves as the Middle East’s foremost AI innovation hubs. Both countries seek to diversify their economies away from oil and gas by strategic investments in AI-driven industries and the development of related infrastructure projects.

AI-Adopting Economies

Emerging economies in Latin America, Africa, and Southeast Asia are developing AI applications while relying on foreign investment, technology transfer, and external regulations to do so. Brazil, Mexico, Argentina, and Chile are building digital ecosystems, but lack strong domestic AI-production capabilities. Brazil excels in fintech, and Chile is advancing its AI agenda through international partnerships (e.g. a binational AI centre with France). Both countries rely on external funding to develop infrastructure and research capacity. Nigeria, Kenya, South Africa, and Egypt demonstrate AI potential in fintech and mobile solutions tailored to local needs. Kenya’s M-Pesa platform exemplifies successful integration of digital financial services, while Nigeria’s expanding start-up ecosystem is promising. However, these countries still face infrastructure, talent, and investment challenges. Southeast Asian countries including Indonesia, the Philippines, Vietnam, Thailand, and Malaysia are also expanding AI applications in digital finance and manufacturing. Indonesia’s growing super-app ecosystem and Vietnam’s emerging technology-manufacturing base represent significant regional developments. Nevertheless, these economies struggle with regulatory gaps, cybersecurity risks, and their dependencies on foreign technology and expertise, particularly in research and development capabilities as well as in semiconductor manufacturing.

Contested AI Zones of Influence

Countries in Eastern Europe, Central Asia, and parts of the Middle East are navigating complex geopolitical tensions while developing their own AI ecosystems. Ukraine, Poland, the Baltic states, and the Balkans carefully balance EU and NATO ties with domestic economic AI priorities. Ukraine is using foreign-supplied defence tech against Russia, Poland is advancing its digital economy infrastructure to attract technology companies (from the US, Japan, South Korea, and Taiwan), and Estonia is leading Baltic initiatives in digital public services and defence technologies. Russia faces challenges in AI development due to restricted semiconductor access and international sanctions. Kazakhstan, Uzbekistan, and Kyrgyzstan are all simultaneously aligned with Chinese, Russian, and Western AI interests, with Kazakhstan and Kyrgyzstan setting up regional AI centres. Uzbekistan, Tajikistan, and Turkmenistan are investing in AI training and innovation (News Central Asia 2025). Meanwhile, in the Middle East, Iran, Israel, and Turkey are using AI in defence and surveillance networks: Israel excels in world-class cybersecurity and defence AI capabilities, Iran is developing AI solutions mostly for its defence sector amid sanctions, and Turkey is integrating AI into its defence industry.

Chips and Checkmates: The Strategic Game of AI Sovereignty

The global divergence among AI strategies has three key consequences: (1) geoeconomic rivalries are intensifying, (2) civil society’s role in AI governance is diminishing, and (3) uncertainties about the future of global AI development are rising. The fragmented AI landscape hinders global standardisation, as the US, China, and the EU increasingly use AI as a tool of geopolitical influence. Simultaneously, Global South countries are emerging as active participants in shaping the international AI order, partly in response to this fragmentation. The Paris Statement (2023) underscores this shift, affirming that states should retain sovereign control over their AI strategies while ensuring alignment with national regulations, which highlights a broader global trend towards digital autonomy and technological self-determination (Élysée 2025). With no multilateral regulation in sight, competition is intensifying as countries seek to control digital resources, secure supply chains, and protect strategic markets. The impact of geoeconomic rivalries is already visible. The US, the EU, Japan, South Korea, Australia, Taiwan, and the Netherlands have taken steps to effectively monopolise AI chip production by restricting the export of advanced semiconductors and manufacturing equipment to China. At the same time, China is investing in domestic chip production to reduce its dependence on Western suppliers. Global South countries, particularly India and Brazil, aim to disrupt this duopoly by developing local AI industries and attracting foreign investment in semiconductor fabrication. The EU launched its ambitious European Chips Act to reduce dependency on Asian and American suppliers.

Policymakers worldwide are employing trade and investment strategies to protect or establish their position in the global AI competition, seeking greater control over key technologies, even at the risk of disrupting international cooperation and markets. This dynamic is evident in the approach of the US towards Ukraine and the Democratic Republic of the Congo (DRC), with the Trump administration prioritising his country’s strategic interest in access to critical minerals essential for advanced technology over other foreign policy considerations.

Both the signatories to the Paris Statement and BRICS advocate for a multipolar AI ecosystem to enhance digital sovereignty. However, their implementation strategies differ: the Paris Statement emphasises national control, while BRICS promotes international cooperation and decentralisation of AI capabilities among its members. Currently comprising ten states, BRICS notably includes India and China – despite their historical tensions – in a shared pursuit of global influence. The bloc’s drive for autonomy aligns with the broader global economic reorientation towards Asia. China’s Digital Belt and Road Initiative functions as a strategic mechanism for its technological cooperation across the Middle East and North Africa. In fact, throughout the MENA region, governments have strengthened collaboration with China to diversify technological ties beyond the West (Belhaj 2024).

Nonetheless, the lack of multilateral regulation and the growing major-power competition in the field of AI are accelerating disparities in technological capabilities. As AI advancements first benefit countries with strong digital infrastructures, skilled workforces, and supportive frameworks, concerns about a global AI divide are intensifying. Indeed, tech corporations and major global powers exploit labour and resources from resource-rich but economically disadvantaged countries (e.g. DRC and Bolivia), reinforcing patterns of “digital colonialism” (Dachwitz and Hilbig 2025). Whereas countries offering favourable regulations and financial incentives attract future technological developments, many low-and-middle-income countries lack the infrastructure for their own AI development (Ray 2025). Leading tech powers are monopolising resources and talent. Meanwhile, countries with flexible AI regulations, such as Singapore and the UAE, attract global talent and corporate investments (Harnoss et al. 2024). Singapore, for instance, is a preferred destination for Chinese AI start-ups seeking to circumvent US trade sanctions, while India focuses on public-sector innovation and digital inclusion. The tech rivalry goes beyond economic competition, reshaping diplomacy, security alliances, and development partnerships. Without coordinated international efforts to address the growing technological divide, the global community risks creating a new hierarchy of “AI-empowered” and “AI-marginalised” countries, undermining inclusive global development and equitable participation in the AI-powered economy.

Voices Unheard: Civil Society Organisations’ Uphill Battle in Shaping AI Governance

As AI is increasingly framed as a national security concern, governments and corporations form alliances prioritising both strategic advantages vis-à-vis rival powers and profit maximisation over public accountability. As a result, civil society organisations (CSO) that highlight the risks of AI misuse, exploitative supply chains, and environmental impacts are often overlooked by policymakers and industry representatives alike.

CSOs play three major roles in AI governance: they are information seekers and providers in a complex matter hard to fully comprehend for both the general public and policymakers, they are advocates for citizens’ rights, and they are watchdogs holding governments and major technology companies accountable. Their importance is recognised by the United Nations’ Internet Governance Forum, which incentivises and encourages civic participation. In the EU, Asia, Africa, and Latin America, several CSOs are pushing for stronger AI oversight. Influential CSOs involved in AI regulations and reforms include AlgorithmWatch, Algorithm Justice League (AJL), Access Now, Avaaz, European Digital Rights (EDRi), the International Association of Privacy Professionals (IAPP), Tactical Tech, and others from across the world. In Europe, organisations including EDRi advocate for enforcing the AI Act, while in the Global South, groups including the African Digital Rights Network call for equitable AI policies, focusing on data sovereignty and environmental impacts.

Before the Paris AI Summit, over 40 CSOs signed an open letter dated 6 February 2025 addressed to the organisers, calling for formal mechanisms to integrate their perspectives into the summit and for sustainable funding to support their involvement. During the summit, CSOs advocated for participatory and democratic AI governance. As a result of these efforts, the Public Interest AI Platform was launched at the Paris AI Summit, fostering collaboration between governments, civil society, and the private sector for responsible AI development, prioritising sustainability, inclusion, and human rights. After the summit, CSOs including Access Now and AJL pushed for stricter AI regulations, including mandatory algorithmic impact assessments and public disclosure of AI-related risks. Moreover, CSOs have increased their fundraising efforts to sustain their oversight and advocacy functions in global AI governance and to conduct public consultations to reinforce the importance of embedding diverse societal voices in shaping global AI reform.

These activities notwithstanding, grassroots initiatives for ethical AI are largely ignored by policymakers and tech corporations, risking the shaping of AI systems by powerful economic and state interests without adequate safeguards for public welfare, human rights, or environmental sustainability. Especially in tech-driven states such as China and the UAE, citizens show higher acceptance of state-led AI adoption due to carefully crafted government narratives around technological progress. This further weakens civil society’s collective influence on AI policies in different governance contexts.

Power Shifts and Paradigm Breaks: The Uncertain Future of AI and the Role of the EU

The acceleration of advances in AI is creating global uncertainties with real risks, emanating from issues such as mis-/disinformation, cybersecurity vulnerabilities, data privacy, and threats to digital infrastructure including subsea cables. New risks are also emerging from user-generated AI systems and unequal access to technology, widening already existing digital divides. In fact, recent global investment patterns suggest a trend towards digital monopolisation, further widening the transatlantic divide on tech governance. Advanced generative-AI models, in particular, pose unpredictable risks, even more so for regions with limited tech resilience, while AI’s growing energy demands add to environmental concerns.

The EU has adopted a risk-based regulatory framework for AI, while the US remains sceptical of European approaches, viewing them as a hindrance to innovation. Accordingly, US vice president J. D. Vance’s appearance at the Paris AI Summit signalled the Trump administration’s exclusive focus on maintaining the United States’ AI dominance.

Moreover, alternative AI innovation hubs in Asia and the Middle East impact Europe’s economic competitiveness, potentially causing businesses to relocate to more permissive jurisdictions. Diverging approaches to AI ethics and data governance – the EU’s rights-based frameworks, China’s state-centric model, and the US market-oriented system – are being adapted across various regions, creating regulatory blocs that magnify global divides over data sovereignty, algorithmic accountability, and digital rights protection, thereby fragmenting the international technological landscape and complicating cross-border innovation and cooperation. The marginalisation of civil society in AI governance represents a troubling democratic deficit at a time when robust civic oversight is most needed.

EU economies must strengthen collaboration with global technology developers, including both major companies and countries with advanced capabilities, while deploying sector-specific regulations to enable innovation-testing under appropriate oversight and encouraging joint public–private investments to advance AI solutions for pressing societal challenges. Germany’s new government has signalled promising initiatives for AI governance. It actively backs the EU’s AI Act and is establishing a national advisory centre to coordinate ethical AI implementation across public administration, paying particular attention to security applications. The new German government’s coalition agreement has committed EUR 500 million to strengthen research infrastructure, nurture AI talent, and fund academic positions. In partnership with France, Germany is creating a joint AI research centre that underscores the two countries’ dedication to European leadership in responsible AI development. Since effective regulation can be achieved only at the EU level, Germany’s coalition government should zealously support Brussels’s efforts in this arena.

The EU should promote a human-rights-based approach to AI governance, supported by robust implementation mechanisms such as mandatory impact assessments before deployment and an AI Human Rights Observatory to monitor compliance. To reduce bureaucratic red tape in AI regulation, the EU should adopt a tiered framework that eases compliance for lower-risk AI applications and establishes fast-track approval for time-sensitive, safety-compliant innovations, supported by periodic reviews aligned with technological advancements. Addressing frontier AI risks requires robust pre-market safety testing and strengthened international cooperation to detect and mitigate potentially harmful applications. Complementing these measures, the EU should enhance its AI research capabilities and infrastructure, ensuring it can lead in both innovation and governance while building diversified partnerships that reduce reliance on any single external actor.

Promoting digital inclusion through initiatives such as an “AI for All” programme and knowledge-transfer partnerships with the Global South would strengthen workforce capacity and equitable access across the European Union. Additionally, CSOs must be embedded in the governance ecosystem through structured, multi-stakeholder advisory councils and supported by EU funding to lead public deliberations, ensuring democratic legitimacy and societal representation in AI policymaking.

AI Policy Study Group (2025), Interim Report, AI Strategy Council/AI Policy Study Group, Government of Japan, Cabinet Office, 4 February, accessed 19 February 2025.

Allen, Karen (2024), Artificial Intelligence Regulation in South Africa: Prioritising Human Security, Institute for Security Studies, 16 December, accessed 21 February 2025.

Belhaj, Ferid (2024), Africa and the Middle East: The Shift from Geopolitics to Geoeconomics, Policy Brief 48/24, Policy Center for the New South, accessed 23 February 2025.

Dachwitz, Ingo, and Sven Hilbig (2025), Digitaler Kolonialismus: Wie Tech-Konzerne und Großmächte die Welt unter sich aufteilen, Munich: C.H. Beck.

Élysée, Website of President of France (2025), Statement on Inclusive and Sustainable Artificial Intelligence for People and the Planet, 11 February, accessed 3 March 2025.

Harnoss, Johann, Viacheslav Romanov, Tanya Mondal, and Mounia Mansouri (2024), Top Talent Tracker STEM AI, Boston Consulting Group, accessed 25 January 2025.

Levy Daniel, Maia (2025), Regional Cooperation Crucial for AI Safety and Governance in Latin America, Brookings, 13 February, accessed 23 February 2025.

Lee, Christy (2025), Russia Turns to China to Step Up AI Race Against US, Voice of America, 9 January, accessed 22 February 2025.

Mok, Charkes, and Athena Tong (2025), Japan’s Pragmatic Model for AI Governance, in: The Diplomat, 13 February, accessed 19 February 2025.

Mukherjee, Supantha (2024), AI, Cloud Funding in US, Europe, and Israel to hit $79 bln in 2024, Accel Says, in: Reuters, 16 October, accessed 21 February 2025.

News Central Asia (2025), Artificial Intelligence in Central Asia: Strategies, Initiatives, Prospects, 3 February, accessed 16 April 2025.

Ray, Trisha (2025), At the Paris AI Action Summit, the Global South Rises, Atlantic Council, 13 February, accessed 10 March 2025.

UNCTAD (2024), Contribution by the Russian Federation to the CSTD 2024-2025 Priority Theme on “Diversifying Economies in a World of Accelerated Digitalization”, United Countries Commission on Science and Technology for Development, 22 February, accessed 16 February 2025.